BOX 16.1: DSM-5 HYPERSOMNOLENCE DISORDER CRITERIA

  Self-reported excessive sleepiness (hypersomnolence) despite main sleep period lasting at least 7 hours, with at least one of the following symptoms:

  •   Recurrent naps or lapses into sleep per day

  •   A prolonged main sleep period of more than 9 hours that is unrefreshing

  •   Difficulty being fully awake after abrupt awakening

  •   Occurs at least 3x per week

  •   Accompanied by significant distress/impairment in cognitive, social, or occupational functioning

  •   Not better explained by another sleep disorder, medical disorder, or external substance

  In a 2020 paper, Plante has described the use of several measurements to establish objective hypersomnolence, in addition to the MSLT. These included total sleep time, the psychomotor vigilance test, and/or pupillometry. Combining these old and new measurements, his team has been able to demonstrate “objective hypersomnolence” in twice as many patients at the Wisconsin clinic than with the MSLT alone.11

  This demonstrates how research teams studying IH and hypersomnolence have been gathering groups of patients with characteristics that are not the same, because of lack of agreement about whom to include and which criteria to use. One group of investigators may hold that only people who fit ICSD-3 criteria or who can be verified to sleep for eleven hours per night should be given the IH label, while others, including Plante, have applied a more flexible framework.

  As a consequence of a looser approach to diagnosis, the people in the Wisconsin study were not uniformly the extreme long sleepers studied by some other researchers. Those who went through a MSLT had an average sleep latency of 9.8 minutes—more than eight minutes, the ICSD-3 cutoff. Only a few displayed total sleep times of over eleven hours. Keep that in mind as we examine the results, which are still helpful in understanding the hypersomnolent brain.

  SLOW DOWN FOR SLOW WAVES

  Plante’s study used high-density EEG to look at the patterns of electrical activity the brain generates during sleep. People with hypersomnolence displayed a localized deficiency in the slow brain waves that occur during the deepest part of non-REM sleep.12 Slow waves are considered a marker of restorative sleep, so this finding could explain why people with hypersomnolence are not getting the same benefits as healthy people from sleep and thus feel the need to sleep for longer time periods.

  Slow wave sleep occurs between 10 to 25 percent of total sleep time in young adults and declines with age. The slow waves appear in short bouts and show up more at the beginning of the night, presumably when the brain is catching up on what it needs. They also occur at a higher density after someone has gone without sufficient sleep, so they are part of the brain’s homeostatic response to staying awake for a long time.

  Slow waves on an EEG recording represent neurons within the brain firing in synchronous oscillations, with active and quiet phases coming about once per second. Slow wave oscillations appear to extend across several regions of the brain, not residing in any one part. The oscillations link the thalamus, in the middle of the brain, and the cortex, closer to the surface, but EEGs mainly detect what is happening in the cortex. The influential sleep researchers Giulio Tononi and Chiara Cirelli, also from the University of Wisconsin, have proposed that slow wave sleep allows a nightly recalibration of the brain’s synapses, facilitating the strengthening of memory for events that occurred during the previous day.

  The deficiency in slow wave activity seen in people with hypersomnolence was not apparent in most previous studies of sleep architecture in IH, which had shown that people with IH don’t display more or less REM sleep than healthy controls.13 A few previous studies did show similar results,14 but they generally took information from the standard number of EEGs used in a polysomnogram. High-density EEGs put a considerably larger number of electrodes (256, compared to six) on someone’s scalp, providing more fine-grained spatial information.

  What the additional resolution revealed is that people with hypersomnolence still experience slow wave sleep, but parts of their brains are not participating as much in the synchronized activity. Reductions in slow wave activity appeared on both sides of the brain and were stronger on the left side. The regions of the brain where the differences were the greatest included the somatosensory cortex, where the brain processes the sense of touch, and the supramarginal gyrus. When someone is awake, the supramarginal gyrus has important functions in language processing and interpreting others’ emotions, but it appears to have a role in sleep quality as well.15

  One major point of the Wisconsin study is that people with hypersomnolence displayed the same characteristics in their high-density EEGs whether they reported symptoms of depression or not. It bolsters Plante’s point about the overlap of psychiatric hypersomnolence with IH and suggests a “shared brain abnormality” in persons with hypersomnolence disorder, independent of the presence or absence of depression.

  The localization of the deficiency in slow wave activity around the parietal lobe was distinct from what Wisconsin researchers had previously observed in a subset of people with obstructive sleep apnea who did not display high levels of subjective sleepiness.16 Also, the limited anatomical extent of the difference in slow wave activity may distinguish hypersomnolence disorder from general aging and neurodegenerative disorders, in which slow wave sleep weakens across a larger extent of the brain. On average, participants in the Wisconsin study were in their twenties. More research is necessary on how slow wave activity in IH varies with age and changes over time.

  The Wisconsin study does not answer the question of why people with hypersomnolence have a partial reduction in slow wave activity. However, it does point to potential ways to address the deficiency. One possible option would be devices that can enhance slow wave sleep through acoustic stimulation.17 These devices harness the observation, dating to the earliest days of EEG studies in the 1930s, that sleepers’ brain waves respond to noises in the room—the effect is sometimes called a K-complex. The response to sound may represent the brain’s attempt to get back on track when external stimuli are knocking at the gates of the thalamus.

  Recent research from Wisconsin and elsewhere shows that acoustic stimulation can enhance memory for vocabulary words learned the previous day, if the device’s sounds are timed to nudge the brain’s oscillations at the right moment. These devices, produced by the electronics giant Philips and also the French company Dreem, have now made it to the consumer market. These devices are attractive for IH treatment because they are nonpharmaceutical interventions, in contrast to stimulants or other medications. However, slow wave–promoting devices’ long-term effects remain untested, and their efficacy in IH has not been examined. Plante had originally planned to include such devices in his study; he told me his team was unable to for practical reasons of scope.18

  Enhancing slow wave sleep through pharmaceutical means is possible, but it is unclear how to do it cleanly, as the history of GHB/Xyrem illustrates. As another example, the compound gaboxadol was studied as a sleep aid by Merck and Lundbeck and was reported to increase slow wave sleep and reduce sleepiness in healthy sleep-deprived adults.19 However, the companies stopped research on the drug in 2007 because of side effects including dizziness, headaches, hallucinations, and tachycardia.20

  BARELY AWAKE

  A second set of studies on the biology of IH comes from a research group led by Thien Thanh Dang-Vu in Montreal. Dang-Vu said his interest in hypersomnia had been piqued during his training. He had published several studies of brain imaging and cognition in other sleep disorders such as sleepwalking or REM sleep behavior disorder. “I began to think about—where are the gaps in knowledge?” Dang-Vu said. “We don’t have a clear idea of what’s happening in NT2 and IH, and filling those gaps could be important both for the field and for patients.”

  His lab’s studies have shown that people with IH display a pattern of reduced cerebral blood flow while awake, compared with controls. The pattern resembles what researchers have observed in healthy young people who are in the middle of slow wave sleep.21 For people with IH, it appears that part of the brain is still in a sleep mode when the person is awake. Partly, this study tells us what we already know: for people with IH, sleep is encroaching on them even when they are awake. Transient automatic behavior, or “microsleeps,” has not been explored in experiments like this but may be consistent with this finding.

  The region of the brain most strongly affected by reduced blood flow was the medial prefrontal cortex, which is involved in activities such as decision making, regulating emotions, and memory retrieval. The reduced blood flow appears as a splotch above and behind the eyes (figure 16.1). Overall, the areas that showed reduced blood flow were part of the default mode network, or DMN, a dispersed set of brain regions responsible for internal awareness and consciousness.

  FIGURE 16.1. Brain imaging when people with IH are awake reveals patterns similar to when healthy people are asleep.

  Source: Soufiane Boucetta et al., “Altered Regional Cerebral Blood Flow in Idiopathic Hypersomnia,” Sleep 40 (2017): zsx140.

  The default mode network consists of regions that are active when someone is not doing anything in particular, especially something that requires focused attention. When someone engages in a specific task, such as reading or solving puzzles, activity in the DMN decreases and others perk up. During sleep, the various parts of the DMN interact with one another less; they appear to disconnect from one another. A way to describe the DMN is as a network responsible for daydreaming, or like the machinery that maintains the idle on a car: the level that the engine returns to when the gas pedal is not pressed.

  To be confident about their results, Dang-Vu and his colleagues had to be certain that their study subjects were not actually sleeping, which is a strong possibility when people with IH are put into a scanner for an extended period of time. The type of imaging they used was SPECT (single photon emission computed tomography), with a radioactive probe that is quickly metabolized. The critical time window lasts just a few minutes, while the probe is taken up by brain cells. Technicians monitored participants to check that they weren’t asleep during this time.

  By current criteria, this was an orthodox IH group. Study participants all fell asleep in an average of eight minutes or less on their MSLTs, and all reported habitual sleep durations of eleven hours or more. If someone reported a higher sleepiness score or fell asleep faster on the MSLT, the pattern of decreased blood flow tended to be stronger.

  The pattern of reduced blood flow in IH didn’t resemble that seen in healthy people who were asked to stay up all night, and it also didn’t look what has been observed in narcolepsy type 1.22 Although the specific neurological injury in narcolepsy type 1 is to the hypothalamus, the consequences of that injury and downstream functional changes can be seen elsewhere in the brain. People with narcolepsy type 1 tend to show reduced blood flow at other nodes of the brain in the frontal and temporal lobes; this is not the same as with IH.23

  The differences suggest that the differences in blood flow between IHers and healthy controls are more connected to the trait of hypersomnolence, rather than the acute state of sleepiness. More imaging research on IHers in different states, awake and asleep, as well as comparisons with other groups, such as people with sleep apnea, could flesh this out.

  As with the Wisconsin study, the patterns of reduced blood flow in IH do not say much about a cause. They do resemble the effects of benzodiazepines, although the reductions in blood flow occurring with benzodiazepines extend more broadly across the brain.24 The findings might fit Rye’s somnogen theory, but they do not point specifically to hyperactive GABA inhibitory circuits.

  In another paper published in 2019, Dang-Vu’s team used MRI (magnetic resonance imaging) to examine differences in brain structure and volume in people with IH.25 They also monitored functional connectivity; this is a way to infer whether separate brain regions are interacting, by observing whether their activities rise and fall together. Functional connectivity within parts of the DMN, such as the medial prefrontal cortex, was less in people with IH than in healthy controls. But in contrast to what the researchers expected, some of the same regions that showed decreased blood flow measurements in IH displayed increased cortical thickness and gray matter volume, in comparison to controls. According to Dang-Vu, this may be a result of the chronically sleepy brain compensating for impaired function.

  IMAGING SLEEP DRUNKENNESS

  Many IHers have the symptom of severe sleep inertia in common: difficulty waking up and impaired alertness lasting minutes to hours after waking. A survey of the Hypersomnia Foundation registry found that a majority of respondents with IH endorsed difficulty waking and the need for multiple alarms (79 percent and 69 percent). Severe sleep inertia can also exist independently of IH or excessive daytime sleepiness. The more evocative term “sleep drunkenness” describes an extension of severe sleep inertia. Sleep drunkenness is defined as confusion or clumsiness upon awakening, with impairments of speech, motor control, or cognition. “I was walking into walls, and I felt like I was in a bubble,” one IHer said at a support group meeting. When experienced clinicians have examined groups of people with IH, they have found that 21 percent to 55 percent of them have sleep drunkenness.26 Yet to qualify as sleep drunkenness, duration and severity of symptoms have not been standardized.

  Sleep inertia and sleep drunkenness can be some of the most disabling aspects of IH because they contribute directly to absences at school or work. The IHer has to depend on someone else to rouse them out of bed and cajole or force them to take their medicine. There may be some overlap with autonomic nervous dysfunction, such that dizziness upon standing induces someone to lie down, facilitating falling asleep again.

  Several medications have been reported to alleviate sleep inertia, including antidepressants, nicotine patches, flumazenil, and gamma-hydroxybutyrate. Taken at bedtime, delayed-release methylphenidate or delayed-release bupropion resolved severe sleep inertia for nineteen out of twenty-two patients, according to the Minnesota sleep specialist Carlos Schenck.27 While some of these medications, such as methylphenidate and bupropion, may be inexpensive and familiar to clinicians, controlled studies focusing on sleep inertia and sleep drunkenness in IH have not been performed until recently.

  Through experiments with healthy volunteers, researchers have learned a lot about sleep inertia, which tends to be greater when someone is awakened out of slow wave sleep or when their body temperature is at its lowest, in the middle of the night.28 Since it is more intense after an interruption of recovery sleep, sleep inertia appears to correlate with slow wave density.29 Sleep inertia may result from the brain waking up piece by piece. Upon waking, blood flow increases first around the brainstem and midbrain, with the anterior cortex, a major element of the default mode network, coming online later.30

  An unresolved question is whether sleep drunkenness, as experienced by people with IH, has the same physiological and neurological characteristics as sleep inertia in healthy people. Bedřich Roth’s work pointed more toward sleep drunkenness being distinct in the setting of IH. He thought sleep drunkenness was more dependent on the patient’s “pathological disposition” rather than on how much time they had been sleeping or their previous stage of sleep.31 In one paper, Roth and his colleagues tried with several healthy control subjects to provoke sleep drunkenness, but they managed to do so only once, by having that person stay up all night beforehand. Through neurological examinations, he observed that sleep drunkenness among IHers included impaired motor coordination, diminished reflexes, and vestibular disturbance. He suggested that the cerebellum may be late to reengage with the rest of the brain upon waking.

  When Lynn Marie Trotti was beginning a brain imaging study at Emory, she said: “We want to find out if sleep drunkenness in IH is the same as what happens to healthy people with sleep inertia and is more pronounced, or whether it’s something different.” In designing her study, Trotti wanted to catch people with IH in the fuzzy state just after they wake up. Participants were asked to take a nap in a designated room near the MRI scanner and upon waking were wheeled into the scanner. To gauge functional impairment, they were given a test of working memory called “N-back,” in which they are supposed to recognize repeated numbers when the numbers are presented one at a time. Afterward, they could relax and were not given any test. Brain imaging requires the subject to stay still, so simultaneous tests of coordination or balance were not possible.

  The study called for participants to go off whatever wake-promoting medication they were on. One participant, a member of the Atlanta hypersomnia support group, reported that this requirement presented some challenges. “I was NOT supposed to fall asleep inside the scanner, which really worried me,” this person told me by email. “The only reason I was able to stay awake for that long was because I had to do something, and then I made up songs to the noises that the machine was making. It was extremely difficult, but I did manage to stay awake.”

  The first paper to be completed from Trotti’s study reported findings similar to Dang-Vu’s.32 This paper used a different type of imaging, PET (positron emission tomography), which measures metabolic activity by looking at how much radioactive glucose brain cells take up. Again, the patterns of metabolic activity were different between IH and narcolepsy type 1. Trotti and her colleagues were able to see several clusters in parts of the default mode network where there was more metabolic activity in IH than in controls. This suggested that IHers’ brains may be working harder to stay awake—echoing a previous study from France. In an earlier PET imaging study of IH, Montpellier researchers had detected only hypermetabolism during the awake state. The authors interpreted this as indicating that IH patients’ brains were working harder to stay awake.33